This work considers the problem of identifying the fault in a faulty dynamical system on the basis of the system's input and output signals only. For this purpose, a model-based method for the design of diagnostic tests which consist of specific input signals and appropriate residual generators is developed. The method extends the structure graph of dynamical systems in order to represent the couplings in a system which has been brought to a specific operating region. The resulting local structure graph is used to determine specific residual generators which can distinguish between faults on the basis of the system's input and output signals in the corresponding operating region. Algorithms to determine advantageous operating regions and input signals which drive the system into such operating regions are given. The application of the method to determine diagnostic tests is demonstrated using a typical automotive system, a throttle valve.

The “HPI Future SOC Lab” is a cooperation of the Hasso-Plattner-Institut (HPI) and industrial partners. Its mission is to enable and promote exchange and interaction between the research community and the industrial partners. The HPI Future SOC Lab provides researchers with free of charge access to a complete infrastructure of state of the art hard- and software. This infrastructure includes components, which might be too expensive for an ordinary research environment, such as servers with up to 64 cores. The offerings address researchers particularly from but not limited to the areas of computer science and business information systems. Main areas of research include cloud computing, parallelization, and In-Memory technologies. This technical report presents results of research projects executed in 2013. Selected projects have presented their results on April 10th and September 24th 2013 at the Future SOC Lab Day events.

In the last decade, AI firmly settled into our industrial society with the expert systems as the representative product. However, almost every one of the systems could cover only a single task domain. In the highly mechanized world of the 21st century, systems will become smart and user friendly enough to cover a wide range of task domains. Systems with much user friendliness must be multilingual because users in different domains usually have different languages. Language is formed in its own culture. Therefore, promotion for cross-cultural scientific interchange will be indispensable for the progress of AI.

Fundamentals of Evaluation and Diagnostics of Welded Structures provides an essential guide to the key principles and problems involved in the analysis of welded structures. Chapter one discusses design issues, key equations and calculations, and the effects of varied heat sources in relation to the temperature field in welding. Chapter two goes on to explore welding stresses and strains. Fracture mechanics and the load-carrying capacity of welded structures are the focus of chapter three. Chapter four considers diagnostics and prediction of the residual life of welded structures, whilst acoustic emission techniques for the analysis of welded structures are reviewed in chapter five. Finally, chapter six supplies supplementary information on numerical techniques and other tests for welded structures.With its distinguished author and detailed coverage, Fundamentals of evaluation and diagnostics of welded structures is an indispensable guide for welding and structural engineers as well as those researching this important topic.

Graph-structured data is ubiquitous throughout the natural and social sciences, from telecommunication networks to quantum chemistry. Building relational inductive biases into deep learning architectures is crucial for creating systems that can learn, reason, and generalize from this kind of data. Recent years have seen a surge in research on graph representation learning, including techniques for deep graph embeddings, generalizations of convolutional neural networks to graph-structured data, and neural message-passing approaches inspired by belief propagation. These advances in graph representation learning have led to new state-of-the-art results in numerous domains, including chemical synthesis, 3D vision, recommender systems, question answering, and social network analysis. This book provides a synthesis and overview of graph representation learning. It begins with a discussion of the goals of graph representation learning as well as key methodological foundations in graph theory and network analysis. Following this, the book introduces and reviews methods for learning node embeddings, including random-walk-based methods and applications to knowledge graphs. It then provides a technical synthesis and introduction to the highly successful graph neural network (GNN) formalism, which has become a dominant and fast-growing paradigm for deep learning with graph data. The book concludes with a synthesis of recent advancements in deep generative models for graphs—a nascent but quickly growing subset of graph representation learning.